■Future Outlook LeTech <3497> expects an increase in sales and ordinary income of more than 20%, with sales of 2.14 billion yen (+33.8% YoY), operating income of 150 million yen (+7.7% YoY), ordinary income of 100 million yen (+21.7% YoY), and net income of 1.03 billion yen (-11.4% YoY) for the July 2024 term, and has maintained its initial forecast (announced in September 2023).

For the fiscal year ending May 2025, projected revenue is 5,770 million yen, reflecting an increase of 30.5% compared to the previous period, with operating profit expected to rise 1.9% to 675 million yen, ordinary profit increasing 3.5% to 675 million yen, and net income expected to increase 2.6% to 443 million yen. The ongoing period involves a global strategy of proactive pre-investment in a growing market, where revenue growth is high, but each profit will be restrained.

High demand for revenue is expected to continue from the previous period, with an increase in project size, estimating a high growth rate of 30.5% compared to the previous period. The domestic market environment is expected to continue to grow steadily, supported by rising investment needs for equipment used in generative AI, large government subsidies for GPU cloud play, and an increase in AI-related budget requests. At G-Deep Advanced <5885>, the strategy is to expand the solution portfolio for further growth, prepare a higher layer of the solution lineup, and strengthen the partner ecosystem. The mid-term management plan positions the two periods from May 2025 to May 2026 as a time for active investment in people and equipment. Furthermore, due to the impact of an increase in project size, gross margin is expected to decrease. As a result of these factors, operating profit is planned to show a slight increase or remain flat (up 1.9% from the previous period).

In the ongoing period, driven by increased demand linked to generative AI and subsidies, there is an unusually high amount of backlog (starting at 2,106 million yen), leading the company to believe that high top-line growth is highly certain. Moreover, past issues with semiconductor supply have been resolved, reducing supply chain risks. Attention will focus on whether adequate investment can be made in internal infrastructure (personnel, equipment, etc.) for the future.

■ Growth Strategy and Topics

Aim to achieve 1 billion yen in operating profit in the final year of the mid-term management plan. Formulate a growth strategy that responds to the era of large-scale AI and cloud transformation.

1. Mid-term Management Plan - Numerical Targets

The company is promoting a mid-term management plan starting from the fiscal year ending May 2025, with the final year ending in May 2027. The target for revenue in the final year is set at 8,415 million yen, with an operating profit target of 1,055 million yen. The average annual growth rates over the three years are expected to be 23.9% for revenue and 16.8% for operating profit, planning to increase revenue by 1.9 times and operating profit by 1.6 times over three years. The first two periods are positioned as investment periods in personnel and equipment, where although top-line growth accelerates, operating profit growth will be restrained. In the final year, due to these results, operating profit is planned to grow significantly, surpassing 1 billion yen. For the final year, the target operating margin is 12.5%, down from the recent achievement of 15.0% (for the fiscal year ending May 2024), which is attributed to the gradual decrease in gross margin due to the expansion of project size.

2. Mid-term management plan growth strategy.

As a growth strategy, three key initiatives will be promoted.

1) Transition to higher-layer solutions and environmental enhancement

The scale of projects for research purposes of generative AI and multimodal AI (AI that handles different types of information such as images and audio) is expanding. The company is expanding its product and service portfolio to higher layers, from desk-side AI workstations to large-scale GPU cloud, to respond timely to the user needs of large-scale AI and promote solid upselling. Additionally, partnerships with domestic datacenters are being formed to provide facilities for stable operation of high-load latest GPU systems, as well as software tools for efficiently training AI, strengthening efforts to comprehensively support efficient AI learning.

2) Strengthening the ecosystem to align with the era of large-scale AI

Domestically, collaboration with SIers has begun to establish a system capable of horizontal and vertical division of labor for large-scale projects. In addition, partnerships with cloud vendors and datacenters are being formed to promote hybrid use of on-premise and cloud. Furthermore, the mobility-focused PoC (Proof of Concept: verifying the feasibility of new ideas, technologies, or methods) center GAT will be further strengthened to develop high-quality projects. Globally, in addition to traditional global partners, the partner network will be further enhanced with OEM vendors, storage vendors, and software vendors closely tied to NVIDIA, enabling procurement strength in products and the construction of original large-scale systems.

3) Promoting hybridization as a comprehensive AI resource vendor

Existing resources on hand will be used for PoC purposes, the latest equipment will be utilized as a private cloud in the datacenter, and a system will be established that accommodates customers' three stages where they burst into public large-scale systems during high-scale computations. Aim to become a comprehensive vendor of AI resources that seamlessly transitions all environments and supports the acceleration of AI development.

3. Initiating the provision of a trial environment for NVIDIA technology aimed at the mobility industry.

Nexity Electronics, an electronics trading company within the Toyota Tsusho Group, has launched a service called "GPU Advanced Test drive" (GAT), which provides a proof of concept environment for users to try out production-grade AI applications, including generative AI, to further accelerate AI development using NVIDIA accelerated computing in Japan's mobility industry starting in April 2024.

GAT not only includes NVIDIA hardware but also a verification environment for NVIDIA's authorized software stack, including NVIDIA AI Enterprise. By utilizing the latest AI development verification environment from GAT, users can make investment decisions quickly. It can be safely used with bare metal physical servers and highly secure lines without using a virtual environment.

With the arrival of the software-defined vehicle era, where value, performance, and evolution are defined by software, the importance and value of software-driven development in the mobility industry are rapidly increasing. Applications utilizing LLM, such as self-driving cars, in-car entertainment, and battery management, are not only being used in service aspects but also seeing a rapid increase in practical applications in the development and manufacturing processes of mobility products, including automobiles.

(Written by FISCO Guest Analyst, Hideo Kakuta)